The Effectiveness of Data Augmentation in Image Classification using Deep Learning
arXiv:1712.04621
Abstract
In this paper, we explore and compare multiple solutions to the problem of data augmentation in image classification. Previous work has demonstrated the effectiveness of data augmentation through simple techniques, such as cropping, rotating, and flipping input images. We artificially constrain our access to data to a small subset of the ImageNet dataset, and compare each data augmentation technique in turn. One of the more successful data augmentations strategies is the traditional transformations mentioned above. We also experiment with GANs to generate images of different styles. Finally, we propose a method to allow a neural net to learn augmentations that best improve the classifier, which we call neural augmentation. We discuss the successes and shortcomings of this method on various datasets.
8 pages, 12 figures
References in corpus (1)
Cited by in corpus (20)
- Further advantages of data augmentation on convolutional neural networks
- Improving Malaria Parasite Detection from Red Blood Cell using Deep Convolutional Neural Networks
- Theoretical evidence for adversarial robustness through randomization
- Augmentation Scheme for Dealing with Imbalanced Network Traffic Classification Using Deep Learning
- Technical Considerations for Semantic Segmentation in MRI using Convolutional Neural Networks
- Real-time Federated Evolutionary Neural Architecture Search
- Exploring Bias in GAN-based Data Augmentation for Small Samples
- Building Robust Deep Neural Networks for Road Sign Detection
- Data Augmentation for Instrument Classification Robust to Audio Effects
- A Meta Approach to Defend Noisy Labels by the Manifold Regularizer PSDR
- Deep learning in bioinformatics: introduction, application, and perspective in big data era
- Snore-GANs: Improving Automatic Snore Sound Classification with Synthesized Data
- Style transfer-based image synthesis as an efficient regularization technique in deep learning
- Learning Optimal Data Augmentation Policies via Bayesian Optimization for Image Classification Tasks
- Thinking Outside the Pool: Active Training Image Creation for Relative Attributes
- Simulation of virtual cohorts increases predictive accuracy of cognitive decline in MCI subjects
- Improved visible to IR image transformation using synthetic data augmentation with cycle-consistent adversarial networks
- Void region segmentation in ball grid array using u-net approach and synthetic data
- Handcrafted vs Deep Learning Classification for Scalable Video QoE Modeling
- Sinusoidal wave generating network based on adversarial learning and its application: synthesizing frog sounds for data augmentation